Aschenbrenner cross-read — he's the demand case, not the nuclear pick (gas-first; zero nuclear in his 13F)

Update

Aschenbrenner cross-read — he's the demand case, not the nuclear pick (gas-first; zero nuclear in his 13F)

  • Type: maintenance (subtype: cross-perspective-evidence)
  • event_id: 2026-06-06-aschenbrenner-cross-read-demand-not-nuclear-pick
  • Source: Aschenbrenner Situational Awareness: The Decade Ahead (June 2024), "Power" section (~p.83–88) — full text 2024-06-04-situational-awareness-the-decade-ahead-full-essay-archived; corrected Q1 2026 13F read 2026-05-20-r19-aschenbrenner-q1-corrected-all-action; full reconciliation dossier 2026-06-05-aschenbrenner-research-dossier.md.
    • parent_event: 2024-06-04-situational-awareness-the-decade-ahead-full-essay-archived
  • Shift: Aschenbrenner — the spine of the parent ai-power-bottleneck perspective — is a power-DEMAND maximalist but a nuclear-SUPPLY skeptic. His June 2024 essay sizes the 100GW / ~20%-of-US-grid shortfall that creates the nuclear TAM, but explicitly routes AROUND nuclear ("building a new gigawatt-class nuclear power plant takes a decade") to natural gas as the near-term answer (Marcellus/Utica shale, ~$100B for 100GW, combined-cycle in ~2yr); SMR appears once, demoted to a deregulation-afterthought basket alongside solar/batteries/geothermal. His corrected Q1 2026 13F holds ZERO nuclear names in any layer (no LEU/CCJ/UEC/UUUU/CEG/VST/TLN/OKLO/SMR/NNE/BWXT).
  • Effect on thesis: No change to the nuclear-fuel-cycle thesis direction. CONFIRMS the conviction gradient (avoid pre-revenue SMR devs — Leo bundles SMR with solar/batteries, never a primary pick) and the AI-driven-demand framing (he IS the demand engine the lane borrows). SHARPENS the anti-thesis / What-Could-Go-Wrong: the foundational AI-power thinker picks gas over nuclear on timeline, and the fund whose book IS the parent perspective holds zero nuclear — a smart-money negative-confirmation worth naming. The lane should answer "why nuclear over gas?" head-on (timeline + the 2024→2026 AI-nuclear re-rate that post-dates the essay; price-check that re-rate via the summaries, do not assert returns here).
  • Per-ticker: OKLO ↓-bias. SMR ↓-bias. NNE ↓-bias (Leo's SMR-as-dereg-afterthought independently confirms "don't size pre-rev SMR as the thesis"). LEU/CCJ/CEG/VST/TLN/BWXT — no new vote (he holds none; absence noted, not a knock on the names themselves).
7 events
DateTypeEvent
Jun 5 InvestigationAschenbrenner research dossier — essay-mined lanes, the pair-trade put book, 13F track-record calibration, and cross-manager convergence May 29 ScanAI Scan May 29 ScanAI Infrastructure Scan Jun 16 ScanAI Scan Jun 20 ScanAI Scan Aug 13 ScanWashout Turn Scan Jun 25 ScanAI Scan

No direct external sources are attached to this read.